“Fat Base” Nucleotide Inhibitor of IMP Dehydrogenase - American

A Potent “Fat Base” Nucleotide Inhibitor of IMP Dehydrogenase†. Wen Wang‡ and Lizbeth Hedstrom*. Department of Biochemistry, Brandeis UniVersi...
0 downloads 0 Views 104KB Size
Biochemistry 1998, 37, 11949-11952

11949

A Potent “Fat Base” Nucleotide Inhibitor of IMP Dehydrogenase† Wen Wang‡ and Lizbeth Hedstrom* Department of Biochemistry, Brandeis UniVersity, Waltham, Massachusetts 02454 ReceiVed May 14, 1998; ReVised Manuscript ReceiVed June 24, 1998

ABSTRACT: Inosine monophosphate dehydrogenase (IMPDH) is a target for anticancer, antiviral, immunosuppressive, and antimicrobial chemotherapy. Thus, IMPDH inhibitors have great potential as chemotherapeutic agents. Here we show that imidazo[4,5-e][1,4]diazapine nucleotide (I) is a potent inhibitor of both human type II and Escherichia coli IMPDH. I is a slow-binding inhibitor. The values of Kd are 1.4 nM and 53 nM for human and E. coli IMPDH, respectively. Inhibition is reversible, as demonstrated by the recovery of activity upon denaturation and renaturation of the enzyme‚I complex. I is not a substrate for IMPDH. I may form a covalent adduct with the active-site Cys of IMPDH. Such an adduct would serve as an analogue for an intermediate in the IMPDH reaction.

Inosine 5′-monophosphate (IMP)1 catalyzes the oxidation of IMP to xanthosine 5′-monophosphate (XMP) with the concomitant reduction of NAD+ to NADH (Figure 1). This reaction is the first committed and rate-limiting step in guanine nucleotide biosynthesis (1). IMPDH inhibitors have antiproliferative activity and have been used as immunosuppressive, antiviral, and anticancer agents (2-5). In addition, IMPDH is believed to be a target useful in the development of antimicrobial chemotherapy (6-8). The mechanism of the IMPDH reaction is shown in Figure 1. The active-site Cys attacks the 2-position of IMP, yielding E-IMP*. This intermediate collapses with expulsion of hydride to NAD+, producing NADH and E-XMP*. E-XMP* hydrolyzes to produce XMP, probably via the tetrahedral intermediate E-XMP†. Although E-XMP* has been isolated and characterized (9-11), the intermediates E-IMP* and E-XMP† have not yet been identified and are most probably unstable outside the IMPDH active site. Coformycin, a potent inhibitor of adenosine deaminase, acts as a transition-state analogue by mimicking the tetrahedral adduct formed during the adenosine deaminase reaction (Figure 2) (12). Similarly, we believe that imidazo[4,5-e][1,4]diazapine nucleotide (“fat base” nucleotide I) and related compounds could be potent inhibitors of IMPDH. Carbonolamines such as I are in rapid equilibrium with the dehydro form; however, the equilibrium concentration of the dehydro form is quite small (Keq ≈ 10-6-10-8) (13-16). The dehydro form of I could react with IMPDH to form a † Supported by the Lucille P. Markey Charitable Trust to Brandeis University. L.H. is a Searle Scholar and a Beckman Young Investigator. * To whom correspondence should be addressed. ‡ Current address: Versicor Inc., 34790 Ardentech Court, Fremont, CA 94555. 1 Abbreviations: IMPDH, inosine 5′-monophosphate dehydrogenase; IMP, inosine 5′-monophosphate; NAD+, nicotinamide adenine dinucleotide; NADH, reduced nicotinamide adenine dinucleotide; XMP, xanthosine 5′-monophosphate; EICARMP, 5-ethynyl-1-β-D-ribofuranosylimidazole-4-carboxamide 5′-monophosphate; DTT, dithiothreitol; ESI, electrospray ionization.

tetrahedral adduct resembling E-IMP* and E-XMP† (Figure 2). Such an adduct would be stabilized by the same interactions that stabilize E-IMP* and E-XMP†, and would be a transition-state analogue of the IMPDH reaction. The following experiments demonstrate that I is a potent, timedependent inhibitor of IMPDH. MATERIALS AND METHODS Materials. IMP, NAD+, and XMP were purchased from Sigma Chemical Co. (St. Louis, MO). [8-14C]IMP was purchased from Moravek Biochemicals (Brea, CA). Human IMPDH type II and Escherichia coli IMPDH were purified as previously described (17). Synthesis of I. The nucleoside of I was synthesized according to published procedures (18). The N-1 position of inosine was alkylated with bromoacetaldehyde diethylacetal. Subsequent ring opening in NaOH, followed by acidification, yielded the nucleoside of I. The identity of the nucleoside was confirmed by NMR and mass spectrometry: calculated molecular weight ) 300 g/mol; electrospray ionization (ESI)-positive ion, m/e ) 301 g/mol; ESInegative ion, m/e ) 299 g/mol. The nucleoside was enzymatically phosphorylated by using Serratia maracens 5′-nucleoside phosphotransferase and p-nitrophenyl phosphate (19). The reaction mixture was extracted twice with ether and adjusted to pH 8 with 1 M NaOH. The mixture was centrifuged and the supernatant was applied to a preparative cellulose thin-layer chromatography plate. The plate was developed in n-propanol:NH4OH:H2O (6:3:1 by vol). I was recovered by removing the cellulose and washing with H2O. The eluant was filtered and absorbed onto a Dionex column, and I was eluted from the column with a gradient of NH4HCO3 (50-1000 mM, pH 8.2). The NH4HCO3 was removed by lyophilization. The identity of I was confirmed by NMR spectroscopy: 1H NMR (D2O) 3.34 (dd, 1 H, C-6 H), 3.54 (dd, 1 H, C-6 H), 4.01 (m, 2 H, C-5′ H), 4.31 (dd, 1 H, C-4′ H), 4.42 (t, 1 H, C-3′ H), 4.65 (t, 1 H, C-2′ H), 5.52 and 5.62 (d, 1 H, C-5 H), 5.68 (d, 1 H, C-1′

S0006-2960(98)01132-5 CCC: $15.00 © 1998 American Chemical Society Published on Web 08/07/1998

11950 Biochemistry, Vol. 37, No. 34, 1998

Wang and Hedstrom

FIGURE 1: Mechanism of the IMPDH reaction.

concentrated to its original volume with a Centricon 30. This dilution-concentration process was repeated 4 times, and the final concentrate was dialyzed overnight against a buffer containing 50 mM Tris and 1 mM DTT. A control sample of E. coli IMPDH (not incubated with I) was treated by the same procedure. Enzymatic activity for both control and treated enzyme was monitored at each step. RESULTS

FIGURE 2: Structures and proposed mechanisms of inhibition of coformycin and I and structure of azepinomycin.

H), 7.51 and 7.50 (s, 1 H, C-1 H). Enzyme Assays. The standard assay solution contains 50 mM Tris-HCl, pH 8.0, 100 mM KCl, 1 mM dithiothreitol (DTT), and 3 mM EDTA. The following substrate concentrations were used: (a) for human IMPDH type II (henceforth human IMPDH) assay, 0.125 mM IMP and 0.100 mM NAD+; (b) for E. coli IMPDH assay, 1.0 mM IMP and 2.5 mM NAD+. The reaction was initiated by addition of enzyme, and the absorbance was recorded at 340 nm for 5 min at 25 °C. Protein concentration was determined by titration with 5-ethynyl-1-β-D-ribofuranosylimdiazole-4-carboxamide 5′-monophosphate (EICARMP) (17). A Hitachi U-2000 spectrophotometer was used for all spectrophotometric analysis. Kinetics of IMPDH Inhibition by I. IMPDH was incubated with various concentrations of I; substrates were added at appropriate times, and the enzyme activity was assayed. The loss of activity followed pseudo-first-order kinetics. InactiVation, Denaturation, and Renaturation of E. coli IMPDH. E. coli IMPDH was incubated with I. After urea was added to a final concentration of 8 M, the solution was incubated for another hour at room temperature. The denatured enzyme was diluted six-fold with a solution containing 50 mM Tris-HCl, 1 mM DTT and 8 M urea, and

Inhibition of IMPDH by I. I was synthesized according to published procedures as described in Materials and Methods. Time-dependent inhibition is observed when both human and E. coli IMPDH are incubated with I (Figure 3). No activity is recovered upon dialysis, although activity is recovered when the E‚I complex is denatured with urea and renatured. This observation indicates that inhibition is reversible. The presence of IMP protects the enzyme against inhibition, whereas the presence of NAD+ has no effect (Figure 3a). High concentrations of IMP completely protect the enzyme from inhibition by I (Figure 3b), which indicates that IMP and I compete for the same binding site. I is not a substrate for IMPDH, however: No NAD+ reduction was observed in the presence of 3.4 µM E. coli IMPDH, 34 µM I, and 1.0 mM NAD+ by an experiment that could detect as little as 1 turnover of NAD+. Determination of Kd for I Inhibition of IMPDH. The values of Kd for I inhibition of both human and E. coli IMPDH were determined by incubating enzyme (6.2 and 5.0 µM of human and E. coli IMPDH, respectively) with IMP (500 µM) and I (5.4 and 11 µM for human and E. coli IMPDH, respectively). The presence of IMP served two purposes: (a) to stabilize the enzyme in the absence of inhibitor; e.g., no activity was lost in the control sample over the course of the experiment (unlike the experiments of Figures 3 and 4, this experiment requires reaction times of days); and (b), as shown above, to compete with I, which increases the observed Kd by the factor (1 + [IMP]/KIMP), where KIMP is the dissociation constant for the enzyme (E)‚ IMP complex. Only E‚I and E‚IMP complexes exist under the conditions of these experiments; no free enzyme will be present. The value of Km for IMP was assigned to KIMP for both enzymes. In the case of human IMPDH, the value of Km for IMP is similar to the value of KIMP determined by isothermal titration calorimetry [Km ) 4 µM versus a dissociation constant ) 1 µM (20, 21)], which justifies this assignment. However, we have not been able to measure KIMP for E. coli IMPDH accurately because the enzyme tends to aggregate at high concentrations. Nevertheless, the results of equilibrium dialysis experiments with the E. coli enzyme are consistent with KIMP ≈ 70 µM, which justifies the assignment of KIMP ) 61 µM from the results of steadystate experiments (22). Activity was monitored over 3.5 days

Fat Nucleotide Inhibitor of IMPDH

Biochemistry, Vol. 37, No. 34, 1998 11951

FIGURE 3: Inhibition of IMPDH by I. (A) Human IMPDH was incubated with I (17 µM) in the absence of substrates (open squares) or in the presence of 125 µM IMP (closed circles) or of 100 µM NAD+ (closed squares) in 100 mM KCl, 1 mM DTT, and 50 mM Tris (pH 7.5, at 25 °C). The open circles denote the enzyme incubated in the absence of I. (B) E. coli IMPDH (0.3 µM) was incubated with I (50 µM) in the presence of various concentrations of IMP. The rate constants for inactivation are determined from plots as in (A). The fraction protected ) 1- k/k0, where k is the rate constant for inactivation in the presence of IMP, and k0 is the rate constant for inactivation in the absence of IMP. Extrapolation of these data to infinite concentrations of IMP by using the equation, fraction protected ) Pm[IMP]/(K + [IMP]), where Pm is the maximal fraction of inactivation and K is the apparent binding constant of IMP, yields Pm ) 0.99 (see solid line). Thus, high concentrations of IMP completely protect IMPDH from inhibition by I, which shows that I is a competitive inhibitor with respect to IMP.

FIGURE 4: Kinetics of I inhibition of IMPDH. (A) Human IMPDH (1.6 µM) was incubated with various concentrations of I: closed circles, no I; open circles, 23 µM I; closed diamonds, 46 µM I; open diamonds, 92 µM I; closed squares, 230 µM I; open squares, 260 µM I; closed triangles, 920 µM I. (B) kobs versus [I] plot for human IMPDH. (C) E. coli IMPDH (0.13 µM) was incubated with various concentrations of I: closed circles, 5.8 µM I; open circles, 14 µM I; closed diamonds, 26 µM I; open diamonds, 35 µM I; closed squares, 46 µM I. No activity is lost when E. coli IMPDH is incubated in the absence of inhibitor (data not shown). (D) kobs versus [I] plot for E. coli IMPDH.

until the reactions reached equilibrium. The equilibrium activities of human and E. coli IMPDH were 24 and 6%, respectively, and the corresponding values of Kd were 1.4 and 53 nM (Table 1). Thus I is a more potent inhibitor of human IMPDH than of E. coli IMPDH; this difference in inhibitor affinity mirrors the difference in IMP affinity. Kinetics of the Inhibition of IMPDH by I. Both human and E. coli IMPDH were incubated with various concentrations of I. The reactions were monitored by assaying activity

at appropriate time intervals. Inhibition followed pseudofirst-order kinetics, as demonstrated by the exponential loss of activity (Figure 4). Replots of the apparent rate constants of inhibition versus concentration of I show biphasic behavior (Figure 4). This observation indicates that the inhibition of IMPDH is a two-step process described by: k1

k2

-1

-2

E + I {\ } E‚I {\ } E‚I* k k

11952 Biochemistry, Vol. 37, No. 34, 1998

Wang and Hedstrom

Table 1. Inhibition of IMPDH by Ia parameter

human

E. coli

Kd (nM) initial Ki (µM) k2 (s-1) kon (M-1 s-1) koff (s-1)

1.4 ( 0. 2 650 ( 50 0.0035 ( 0.0001 5.2 ( 0.2 (7 ( 1) × 10-9

53 ( 6 40 ( 8 0.0025 ( 0.0002 63 ( 6 (3.3 ( 0.5) × 10-6

a Enzyme and various concentrations of I were incubated in a solution of 50 mM Hepes, 100 mM KCl, and 1 mM DTT at 25 °C. Formation of E‚I was monitored by loss of activity (Figure 4A and C). The initial Ki, k2, and kon () k2/Ki) were determined from the plots in Figure 4B and D. The final dissociation constant, Kd, was determined by incubating IMPDH (6.2 and 5.0 µM of human and E. coli IMPDH, respectively), I (5.4 and 11 µM for human and E. coli IMPDH, respectively), and 500 µM IMP. No free enzyme will be present under these conditions. The samples reached equilibrium in 3 days. No activity was lost in control samples containing only enzyme and IMP. The value of Kd for the E‚I complex was calculated according to: Kd ) [E‚IMP][I]free/ [E‚I] (1 + [IMP]/KIMP) where [E‚IMP] and [E‚I] were determined from the remaining activity and [I]free ) [I]0 - [E‚I]. The value of koff was calculated from Kd ) koff/kon. The values of KIMP are 4 and 61 µM for human and E. coli IMPDH, respectively).

The data were fit to the following equation:

kobs ) k2[I]/(Ki + [I]) Assuming that k-1 . k2, which seems reasonable given the magnitude of k2, then Ki ) k-1/k1, kon ) k2/Ki, and koff ) konKd. Since k-1 is a fast step, koff will equal k-2. Table 1 summarizes the kinetic parameters for the inhibition of human and E. coli IMPDH by I. The value of kon for I inhibition of E. coli IMPDH is 100-fold greater than for the human enzyme. More-rapid inhibition appears to be a general property of the E. coli enzyme: EICARMP also inactivates E. coli IMPDH 10-fold faster than it does the human enzyme (17). DISCUSSION Two unstable tetrahedral intermediates, E-IMP* and E-XMP†, are formed during the IMPDH reaction (Figure 1). Compounds that form enzyme adducts resembling these intermediates are expected to be very potent inhibitors of IMPDH. Although I is predominantly hydrated in solution, it is nevertheless in rapid equilibrium with a small concentration of the dehydro form (Figure 2). Precedents for this chemistry are well documented (13-16). The dehydro form reacts with IMPDH to form a tetrahedral adduct that resembles E-IMP* and E-XMP†. This adduct would be a transition-state analogue for IMPDH, and this reaction is analogous to the reaction of aldehyde transition state analogue inhibitors with serine and cysteine proteases. The experiments described above demonstrate that I is a potent inhibitor of both human and E. coli IMPDH, as predicted. Although inhibition may result simply from the association of I with IMPDH, it seems reasonable to believe, based on chemical precedent, that a covalent adduct forms. The E‚I complex is unstable when the enzyme is denatured, an observation generally taken as evidence that the complex is noncovalent; however, the enzyme-I adduct shown in Figure 2 would not be stable outside the active site of the enzyme (again, such behavior is observed in aldehyde inhibition of serine and cysteine proteases). Unfortunately, this observation precludes confirmation of adduct formation by mass spectroscopy, and the similarity of the hydrated I

and the adduct, as well as the 220 kDa molecular mass of the IMPDH tetramer, suggests that NMR experiments using 13 C-labeled I would not be informative. Therefore, further identification of the adduct will require solution of the X-ray crystal structure of the E‚I complex. The inhibition of IMPDH by I is time dependent; k2 and k-2 may represent formation and decomposition of a covalent adduct (Figure 2) or may represent a conformational change. Of course, these alternatives are not mutually exclusive: The covalent intermediate could form and a kinetically significant conformational change could also occur. Similar inhibitors of IMPDH may already exist in nature. The structure of azepinomycin, a natural product with antibiotic and antitumor activity (23), is very similar to that of the base of I (Figure 2). The activity of azepinomycin is believed to result from inhibition of guanase. Our work on I suggests that the nucleotide of azepinomycin will also be a potent inhibitor of IMPDH. Perhaps the antibiotic and antitumor properties of azepinomycin result from the inhibition of IMPDH. REFERENCES 1. Weber, G. (1983) Cancer Res. 43, 3466-3492. 2. Streeter, D., & Koyama, H. (1976) Biochem. Pharm. 25, 2413-2415. 3. Witkowski, J. T., Robins, R. K., Sidwell, R. W., & Bauer, R. J. (1972) J. Med. Chem. 15, 1150. 4. Robins, R. (1982) Nucleosides Nucleotides 1, 35-44. 5. Allison, A. C., Almquist, S. J., Muller, C. D., & Eugui, E. M. (1991) Transplant. Proc. 23, 10-14. 6. Hupe, D., Azzolina, B., & Behrens, N. (1986) J. Biol. Chem. 261, 8363-8369. 7. Verham, R., Meek, T. D., Hedstrom, L., & Wang, C. C. (1987) Mol. Biochem. Parasitol. 24, 1-12. 8. Webster, H. K., & Whaun, J. M. (1982) J. Clin. InVest. 70, 461-469. 9. Huete-Perez, J. A., Wu, J. C., Witby, F. G., & Wang, C. C. (1995) Biochemistry 34, 13889-13894. 10. Link, J. O., & Straub, K. (1996) J. Am. Chem. Soc. 118, 20912092. 11. Fleming, M. A., Chambers, S. P., Connelly, P. R., Nimmesgern, E., Fox, T., Bruzzese, F. J., Hoe, S. T., Fulghum, J. R., Livingston, D. J., Stuver, C. M., Sintchak, M. D., Wilson, K. P., & Thomson, J. A. (1996) Biochemistry 35, 6990-6997. 12. Wolfenden, R., & Kati, W. M. (1991) Acc. Chem. Res. 24, 209-215. 13. La Londe, R. T., & Eckert, T. S. (1983) J. Org. Chem. 48, 3841-3842. 14. Fife, T. H., & Hutchins, J. E. C. (1980) J. Org. Chem. 45, 2099-2104. 15. Cooper, A. J. L., & Redfield, A. G. (1975) J. Biol. Chem. 250, 527-532. 16. Capon, B., & Connett, B. E. (1965) J. Chem. Soc. , 4497. 17. Wang, W., Papov, V. V., Minakawa, N., Matsuda, A., Biemann, K., & Hedstrom, L. (1996) Biochemistry 35, 95101. 18. Aoyagi, M., Minakawa, N., & Matsuda, A. (1991) Symp. Nucleic Acid Chem. 25, 57-58. 19. Fyfe, J. A., Keller, P. M., Furman, P. A., Miller, R. L., & Elion, G. B. (1978) J. Biol. Chem. 253, 8721-8727. 20. Wang, W., & Hedstrom, L. (1997) Biochemistry 36, 84798483. 21. Bruzzese, F. J., & Connelly, P. R. (1997) Biochemistry 36, 10428-10438. 22. Kerr, K. M., & Hedstrom, L. (1997) Biochemistry 36, 1336513373. 23. Umezawa, H., Takeuchi, T., Iinuma, H., Hamada, M., & Nishimura, S. (1983) Jpn. Kokai Tokkyo Koho JP 58,159,494. BI981132W